Inference
Explore
Leaderboard
Benchmarks
Labs
Run locally
Search
⌘K
EN
English
Русский
Локально
中文
本地运行
12 Nov
2026
Benchmarks
/
Humanity's Last Exam
Center for AI Safety et al.
Humanity's Last Exam
In index
Knowledge
Best in knowledge
Source ↗
Models
?
Models with a result.
77
Top result
?
Best result on this benchmark.
74.2
%
Claude Fable 5.2
Top-3 spread
?
Points from first to third.
9.2
pts
Year
?
Year of release.
2025
Result and price
0
20
40
60
80
100
$0.03
$0.1
$0.3
$1
$3
$10
$30
$100
$300
RESULT
PARETO
OUTPUT PRICE PER 1M TOKENS · LOG SCALE
57 of 77 results
Download
PNG
Results
77 results
Export
CSV
Export
JSON
#
Model
Result
?
The score from the source.
01
Claude Fable 5.2
74.2
%
02
MA
Kimi K3.1 Pro
Above fit
?
Far above this model's index.
70.0
%
03
Claude Fable 5.1
65.0
%
04
Claude Opus 5
64.7
%
05
Claude Mythos 5
Above fit
?
Far above this model's index.
64.5
%
06
Muse Spark 1.1
Above fit
?
Far above this model's index.
62.1
%
07
TE
Hy4
59.1
%
08
GPT-5.4 Pro
58.7
%
09
Claude Opus 4.8
57.9
%
10
Claude Sonnet 5
Above fit
?
Far above this model's index.
57.4
%
11
GPT-5.5 Pro
57.2
%
12
AP
Apodex 1.1
Above fit
?
Far above this model's index.
56.1
%
13
MA
Kimi K3
56.0
%
14
TE
Hy4 preview
55.4
%
15
Claude Opus 4.7 (Adaptive)
54.7
%
16
TM
Inkling 2 Small
53.2
%
17
Claude Opus 4.6
Above fit
?
Far above this model's index.
53.0
%
18
DS
dots3-note Preview
52.6
%
19
GPT-5.5
52.2
%
20
GPT-5.4
52.1
%
21
Muse Spark
50.4
%
22
Claude Sonnet 4.6
49.0
%
23
MiMo-V2.5-Pro
48.0
%
24
TM
Inkling-Small
Above fit
?
Far above this model's index.
47.8
%
25
IN
Agents-A1
47.6
%
26
ST
Step 5 Preview
46.5
%
27
TM
Inkling
46.0
%
28
OA
Ornith-1.5-397B
44.6
%
29
Qwen3.8 Max
43.6
%
30
DeepSeek V4 Pro 0813
42.7
%
31
DeepSeek V4.1 Flash 1017
42.6
%
32
GPT-5.4 mini
41.5
%
33
Qwen3.7 Max
41.4
%
34
GLM-5.2
40.5
%
35
Gemini 3.5 Flash
40.2
%
36
GPT-5.4 nano
37.7
%
37
IN
Ling 3.1 Flash
37.4
%
38
Qwen3.8-Omni-Flash
36.5
%
39
Qwen3.8-Flash-Next
35.9
%
40
Grok 4.3
35.0
%
41
DeepSeek V4 Flash 0731
34.8
%
42
MA
Kimi K2.6
34.7
%
42
Qwen3.7 Plus
34.7
%
44
GLM-5.1
31.0
%
45
Qwen3.8-27B
30.8
%
45
Claude Opus 4.5
30.8
%
47
GLM-5
30.5
%
48
MA
Kimi K2.5
30.1
%
49
Nemotron 3.5 Super VL 120B A12B
29.2
%
50
Nemotron 3.5 Super VL 120B A12B NVFP4
29.0
%
51
Qwen3.6 Plus
28.8
%
52
Qwen3.5 397B
28.7
%
53
MiniMax M2.7
28.0
%
54
ST
A.X K2
27.8
%
55
Nemotron 3 Ultra
26.7
%
56
Gemma 4 31B
26.5
%
57
OA
Ornith-1.5-35B-A3B
25.6
%
58
TE
Hy3 Preview
25.5
%
59
GLM-4.7
24.8
%
60
Qwen3.6-27B
24.0
%
61
IN
Ling 3.0 Flash
22.7
%
62
Qwen3.6-35B-A3B
21.4
%
63
OA
Ornith-1.5-9B
20.2
%
64
Gemini 2.5 Pro
18.8
%
65
LA
K-EXAONE 2.0
18.3
%
65
Nemotron 3 Super 100B
18.3
%
67
Qwen3 235B 2507 (Reasoning)
18.2
%
68
DeepSeek-R1-0528
17.7
%
69
Gemma 4 26B A4B
17.2
%
70
DeepSeek V3.1 (Reasoning)
15.9
%
71
GPT-OSS 120B
14.9
%
72
GPT-OSS 20B
10.9
%
73
Nemotron 3 Nano 30B
10.6
%
74
OP
MiniCPM5-2B
8.9
%
75
LI
LFM2.5-24B-A2B
7.4
%
76
LI
LFM3-1.7B
7.2
%
77
MA
Kimi K2
4.7
%
